Vermont Legislative Joint Fiscal Office

One Baldwin Street Montpelier, VT 05633-5701 (802) 828-2295 Fax: (802) 828-2483

ISSUE BRIEF
Date: June 1, 2016
Prepared by: Joyce Manchester

Medicaid Enrollments: A Methodological Comparison

Executive Summary
An estimate of the total number of adult Vermonters who would enroll in some sort of Medicaidrelated health insurance program in fiscal year 2015 was quite close. However, the average number
of adults enrolled in full Medicaid coverage as General Adults or New Adults in fiscal year 2015 was
about 24,000 or 50 percent greater than the estimated number as of January 2014.1 The number
enrolled in full Medicaid coverage in February 2016 was about 19,000 or 29 percent greater than
the estimated number as of January 2015. Over the 2-year span, the difference between the
January 2014 estimate and the actual enrollment in February 2016 was about 37,000.
Not quite half of the difference between the early estimate and the February 2016 actual
enrollment can be explained by two factors:
1. The Affordable Care Act (ACA) imposed new Medicaid eligibility rules in 2014:

The income limit rose for some types of full Medicaid coverage

The definition of income changed; some adults became newly eligible but others became
ineligible

2. With approval from the Centers for Medicare & Medicaid Services (CMS), Vermont initially
implemented the Affordable Care Act (ACA) under the old rules:

A number of adults were automatically enrolled in Medicaid in 2014 using the old income
definition, and no verification of eligibility was conducted for almost 24 months

The original estimate for new enrollments in 2014 also relied on the old income definition

Problems with Vermont Health Connect prohibited eligibility redeterminations for adults, including
those who had a change of circumstance, between January 2014 and late 2015. As a consequence,
Vermont had not verified income eligibility since the new rules became effective. Verifying income is
now underway, and the State should have better information on the number of eligible Medicaid
enrollees by late fall of 2016.
1

Full Medicaid coverage means that Medicaid is the primary source of health coverage. The estimates were made by a
joint administration-legislative staff working group based on the best information available at the time. The goal of this
Brief is to explain the divergence in an effort to better inform future estimates.
VT LEG #318187 v.1

This analysis will focus on applying the old and new eligibility rules to tax data from 2013 and
analyzing the auto-enrollments. By doing so, JFO can explain up to 45 percent of the unexpected
increase in the number of adults enrolled as General Adults or New Adults with full Medicaid
coverage.
Explaining the rest is less clear but could arise from a number of factors. They include the new
mandate for health insurance coverage, a less complex eligibility process under the ACA, increased
outreach and publicity surrounding the ACA, and the possible eligibility of Medicaid adults for whom
no tax data exist because they do not file tax returns.
In an illustrative exercise, JFO finds roughly 71,500 adults potentially eligible for full Medicaid in
2016. That estimate contrasts with actual enrollment of about 84,000 in February 2016.
Estimated vs. Actual Enrollments and a Brief Description of the Analysis
1. Projections in January 2014 of the total number of adults who would enroll in some sort of
Medicaid-related health insurance program in state fiscal year (SFY) 2015, including both full
coverage and Vermont Premium Assistance, were not too far off the mark. The table below shows
the original estimate for SFY 2015 as presented to the E-Board in January 2014 as well as the actual
average enrollment for SFY 2015.
For SFY'15

General Adult
New Adult
Premium Assistance
Sum

Orig. Est
E-Board
Jan'14
11,679
35,059
42,785
89,523

BAA Est
E-Board
Jan'15
15,504
48,500
18,007
82,011

Actual
Average
Enrollment
SFY15
17,381
55,153
16,906
87,440

Actual
Minus
Jan'14
Projected

-2,083

However, the original projection in January 2015 of the total number of adults who would enroll in
some sort of Medicaid-related health insurance program in SFY 2016 was not so close.
For SFY'16

General Adult
New Adult
Premium Assistance
Sum

Orig. Est
E-Board
Jan'15
15,966
48,985
18,368
83,319

BAA Est
E-Board
Jan'16
20,228
58,292
17,244
95,764

Through
1st 8 Months
Feb'16
20,906
63,173
13,242
97,321

Actual
Minus
Jan'15
Projected

14,002

VT LEG #318187 v.1

2. The number of adults with full Medicaid coverage enrolled as General Adults or New Adults in
both SFY 2015 and SFY 2016 surpassed the number expected by quite a wide margin (see Figures 1
and 2 below).

Figure 1. SFY'15 Estimated and Actual

Medicaid Enrollments
General Adult

Total = 46,738

SFY15

New Adult

Total = 70,534
53,153

Difference = 23,796
or 51% higher

35,059
11,679

17,381

Original Estimate E-Board

Actual Average

Figure 2. SFY'16 Estimated and Actual

Medicaid Enrollments
General Adult

Total = 64,951

48,985

New Adult

SFY16

Total = 84,079

Difference = 19,128
or 29% higher

63,173

15,966

20,906

Original Estimate E-Board

Through 1st 8 Months

Over the 2-year period, the number of adults on full Medicaid in February 2016 was 37,341 above
the January 2014 estimate for SFY 2015.2
As seen in Figure 3 below, actual enrollments jumped in January 2014 and have continued to grow.

The January 2014 E-Board estimate was 46,738. Actual enrollment in February 2016 was 84,079. The difference is
37,341.
VT LEG #318187 v.1

Figure 3. Enrollment in VHAP and Medicaid: General Adults and

New Adults, Oct 2005 through Feb 2016
70,000

ACA
begins
Jan 2014

60,000

New Adults
63,173

50,000

Dec 2013
40,000
30,000
20,000

VHAP
21,803

Early est's
General adults
20,906

General Adults
9,315

10,000

Early est's
Oct-05

Oct-06

Oct-07

Oct-08

Oct-09

Oct-10

Oct-11

Oct-12

Oct-13

Oct-14

Oct-15

3. Analyzing the effect of changes in the income rules and automatic enrollments from old
Medicaid to new Medicaid can explain up to about 45 percent of the unexpected increase in
enrollment.3 Using Vermont tax data from 2013 and data on the automatic enrollments from old to
new Medicaid leads us to the following estimates:

Between 9,700 and 11,700 adults might have been automatically transferred from VHAP and
Catamount to new Medicaid based on the old definition of income for those programs but
would not have been eligible based on the new definition of income for Medicaid. As noted
on page 8, Vermont obtained a waiver from CMS allowing the state to auto-enroll individuals
deemed eligible under the old rule methodology so that expenses associated with this
population would be deemed eligible for Medicaid match. Because the redetermination process
is just beginning, large proportions of those adults remain on Medicaid in early 2016.

About 5,200 newly eligible self-employed adults and farmers might have been missed in the early
estimates because the definition of income used to determine eligibility changed under the ACA.

We do not have a complete understanding of the remaining 20,000 to 22,000 adults, but a
number of factors may play a role. Some adults may have applied in response to the individual
mandate for health insurance coverage; some may have been eligible under the old rules but
were not on the rolls in 2013 because they were not up-to-date on paying premiums; some may
have moved from another Medicaid eligibility category; others may have qualified through
3

JFO thanks the Department of Taxation for providing aggregated tax data for 2013.
VT LEG #318187 v.1

Vermont Health Connect on self-attested income information that may not stand up when
verified. In addition, we have no income information on the 8 percent of adult residents ages 23
to 64 who do not file taxes in Vermont, but many of them are likely to be eligible for full Medicaid
coverage. The illustrative exercise in the section below suggests about 70,000 adults ages 18 to
64 might be expected to be eligible for full Medicaid coverage in total in 2016.
Illustrative Estimate. Under the ACA, adults are first assessed for eligibility under the new income
rules that use Modified Adjusted Gross Income (MAGI) as the starting point. If they are found
ineligible, states are required to re-assess them for other eligibility categories using the pre-ACA
income guidelines. Based on analysis of 2013 tax data, about 58,000 tax-filing adults ages 23 to 64
appear eligible for full Medicaid under the post-ACA rules; others may have been eligible under the
pre-ACA income rules. Some of the 58,000 tax-filing adults would have had employer-sponsored
insurance through their workplace. In 2014, about 30 percent of people in U.S. households with
income under $25,000 had private health insurance. Because some with private insurance can
remain eligible for Medicaid, we estimate between 17 percent and 20 percent of those low-income
households in Vermont would not receive Medicaid. We have no data on the almost 29,000 nonfilers but estimate that about 60 percent could be eligible for Medicaid. Moreover, the U.S. Census
Bureau estimates the population of adults ages 23 to 64 is 5,837 smaller in 2016 than it was in 2013;
one-fifth to one-quarter might have qualified for Medicaid. Based on those rough calculations and
using the midpoint for each assumption, roughly 63,000 adults ages 23 to 64 might be expected to be
eligible for Medicaid in 2016.
58,000
- 10,730
47,270

adult tax filers (23-64) who appear eligible for Medicaid in 2013 using post-ACA rules
if 17% to 20% have private health insurance and no Medicaid

+17,380
64,650

if 60% of the non-filers (23-64) were eligible for Medicaid

-1,313
63,337

if 1/4 to 1/5 of the adults no longer in the age group 23-64 would have qualified for Medicaid
Rough estimate of population ages 23-64 eligible for Medicaid in 2016

In addition, a number of Vermonters ages 18 to 22 are eligible for full Medicaid as General Adults or
New Adults. In December 2015, 9,658 Vermonters ages 18 to 22 were enrolled as General Adults and
New Adults. However, two issues arise. First, 2,250 Vermonters enrolled in Medicaid at ages 19-22
are enrolled in the Dr. Dynasaur Child category. Dr. Dynasaur covers children in low-income families
until they turn age 19. Many of those enrollees will qualify for Medicaid in a different category as
they become ages 19 and 20. To continue the illustrative exercise, we assume half of them will do so
and the other half will be found ineligible. Second, the Vermont State Auditor reports that 22 percent

VT LEG #318187 v.1

of Dr. Dynasaur customers were delinquent on premium payments.4 Applying that percentage to the
age 18 Dr. Dynasaur enrollees and removing half of the age 19 to 22 Dr. Dynasaur enrollees leaves
about 8,200 Vermonters ages 18 to 22 in full Medicaid. Adding those younger people to the estimate
above yields a rough estimate of about 71,500 Vermonters ages 18 to 64 who are likely to be eligible
for full Medicaid in 2016. That rough estimate is about 15 percent below current enrollment.
9,658 ages 18-22 enrolled Dec 2015

63,337 estimate for ages 23-64

-1,125 if half Dr D ages 19-21 ineligible

+8,226 estimate for ages 18-22

8,533
-307 if 22% Dr D delinquent, age 18

71,563 total rough estimate

8,226

Rough estimate for ages 18-22

84,079 actual
enrollment
in Feb
actual
enrollment
Feb2016
84,079 2016
85.1% rough estimate vs. actual
83.4% estimate vs actual

A better idea of Medicaid enrollments will come after the current round of redeterminations is
completed. Some of the findings here may inform what to expect from those redeterminations.
The remainder of this memo will explain

How eligibility rules changed for Medicaid and the Vermont health insurance programs

The extent to which 2013 tax data is useful in examining the pre-ACA and post-ACA transition

How information about automatic enrollments in Vermont helps to understand the transition

How did eligibility rules for Medicaid and Vermont health insurance programs change in 2014?
Categorical eligibility for General Adults did not change, but income eligibility rules for Medicaid
changed in two important ways in 2014:

The income limit rose, making more adults eligible, and

The definition of income used to determine eligibility changed, causing some adults to become
eligible but others to become ineligible. Adults who are ineligible under the new income test may
still be eligible under a different eligibility category using the former income rules.

1. The income limit for Medicaid rose from 100 percent of the Federal Poverty Level (FPL) to 138
percent of FPL. Note that FPL depends on the number of persons in the household (see Figure 4).
The federal poverty level thresholds are updated each year by the Census Bureau. Federal poverty
level thresholds for 2013 determine Medicaid eligibility in 2014.
4

Report of the Vermont State Auditor, Vermont Health Connect: Future Improvement Contingent on Successful
System Development Project. April 14, 2015. Available at
http://auditor.vermont.gov/sites/auditor/files/Final%20VHC%20Report%20Repost%206.9.2015.pdf.
VT LEG #318187 v.1

Figure 4. 2013 Income Levels for Medicaid

by Family Size,
Pre-ACA 100% FPL and Post-ACA 138% FPL
100% FPL

138% FPL
$32,499

$31,590

$27,570

$23,550

$19,530

$15,510

$11,490
1

$26,951

$21,404

$15,856

$43,594

$38,047

Number of Persons in Household

2. The definition of income was changed to become less complex; the resource test was eliminated
for people who qualify under the new income definition (modified AGI, or MAGI). Some eligibility
categories retain a resources test. VHAP and Catamount did not have a resource test.
The pre-ACA definition of income for Medicaid and Vermonts health insurance programs started with a broad
measure of incomeincluding veterans benefits, workers compensation, and child support receivedand
then allowed several important deductions. In particular,

earners could deduct up to $1,080 per year as a standard employee expenses deduction (SEED)

earners with dependent care expenses could deduct up to $2,100 per child age two or older per year and
up to $2,400 per child under age 2 per year

The post-ACA definition of income includes all the additions and deductions that are part of adjusted gross
income (AGI). It thereby allows two deductions that can be significant for a number of Vermonters:

business losses and depreciation for self-employed individuals

business losses and depreciation for farmers

The table below summarizes the changes in the definition of income pre- and post-ACA.

VT LEG #318187 v.1

Pre-ACA Income Definition

Post-ACA Income Definition

Wages (1040 line 7)

+ Taxable Interest (1040 line 8a)
+ Ordinary Dividends (1040 line 9a)
+ Social Security Benefits (1040 line 20a)
+ Pensions and Annuities (1040 line 16a)
+ Unemployment Compensation (1040 line 19)
+ Alimony Received (1040 line 11)
+ Self-Employed Income (Schedule C line 7)
+ Farm Income (Schedule F line 9)

AGI (1040 line 37)

+ Non-taxable social security benefits (1040 line
20a-20b)
+ Tax-exempt interest (1040 line 8b)
+ Foreign Earned Income

- First $50 received in child support

- Standard Employee Expenses Deduction
up to $1,080/year.
- Dependent Care Expenses up to $2100/year.
- Certain Self-Employment Deductions
- Certain Farm Income Deductions

- Scholarships, Awards, Fellowships, and Grants

- American Indian Income
- Lump Sums

Indicates data not available and therefore not included in Tax/JFO analysis

3. Vermont health insurance programs pre-ACA generally used the same income definition as preACA Medicaid but higher income cut-offs, except for students. Those higher income cut-offs together
with the changes in income definition are important when looking at the transition from old public
insurance programs to new Medicaid.5 VHAP and Catamount did not have a resource test.
VHAP = Vermont Health Assistance Program

Required no access to employer-sponsored insurance (ESI) and no health insurance within 12

months

Income up to 150% FPL for childless adults; up to 185% FPL for parents or caretakers

VHAP ESIA = VHAP Employer Sponsored Insurance Assistance: premium assistance

Must be eligible for VHAP and have access to employer-sponsored insurance

CHAP = Catamount Health Assistance Program: premium assistance

Must be uninsured for 12 months with no access to ESI or have a qualifying event

Must not be eligible for VHAP and have income at or below 300% FPL; if income between 200%
and 300% of FPL, disregard $400 in earned income per household

Source: http://humanservices.vermont.gov/on-line-rules/esd/premium-assistance-5900/view
VT LEG #318187 v.1

Catamount ESIA = Employer Sponsored Insurance Assistance: premium assistance

Must be uninsured for 12 months with access to ESI, or have a qualifying event

Must not be eligible for VHAP, income at or below 300% FPL; if income between 200% and 300%
of FPL, disregard $400 in earned income per household

The old higher income thresholds suggest that some adults who were on VHAP and many adults
who were on Catamount might not be eligible for new Medicaid, but only careful analysis can put
all the pieces together.
What can be learned from tax data regarding enrollments from 2013 to 2014?
We can use data on Vermont tax filers from 2013 to look at tax filers only; tax filers represent about
92 percent of Vermont adults ages 23 to 64.
1. The availability of the standard employee expenses deduction (SEED) pre-ACA made a significant
difference to many people. Most Vermont incomes are higher using the post-ACA income definition,
implying that fewer people would qualify for Medicaid, contrary to expectation when thinking about
moving from 100 percent of FPL to 138 percent of FPL.
Under the pre-ACA income definition, anyone who was working was eligible for the standard
employee expenses deduction (SEED). It was $90/month or up to $1,080/year, but monthly income
is not known. The analysis of tax data used the lesser of wages or $1,080, so most people got the
whole deduction. That $1,080 seems to trump all the other adjustments included in the post-ACA
income definition.
When the $1,080 SEED is included in pre-ACA incomes, 81 percent of households with pre-ACA
incomes below 300 percent of FPL who filed taxes in 2013 had higher post-ACA incomes compared to
pre-ACA incomes (see Figure below). 19 percent of households had lower post-ACA incomes and 1
percent of households had about the same income.
When the $1,080 SEED is omitted, 20 percent of households with pre-ACA incomes below 300
percent of FPL who filed taxes in 2013 had higher post-ACA incomes compared to pre-ACA incomes.
29 percent of households had lower post-ACA incomes; 51 percent of households had about the
same income.

VT LEG #318187 v.1

2. The availability of the deductions for business losses and depreciation for the self-employed and
farmers makes a difference as well.
Of the filers who could be called self-employed (filed a schedule C), about 7,000 adults would have
qualified for Medicaid using the post-ACA income definition but not for Medicaid or VHAP using the
pre-ACA definition. The reverse is true for about 2,500 adults. The net gain is up to 4,500 selfemployed. Some may have qualified for VHAP/Medicaid earlier, but we have no information on their
deductions related to being self-employed.
Of the filers who could be called farmers (filed a schedule F), about 800 adults would have qualified
for Medicaid using the post-ACA income definition but not for Medicaid or VHAP using the pre-ACA
definition. The reverse is true for about 100 people. The net gain is up to 700 farmers. Again, some
may have qualified for VHAP/Medicaid earlier, but we have no information on their farm business
deductions.
Taken together, newly available business losses for the self-employed and farmers could have added
about 5,200 adults on net to the Medicaid enrollments. Those additions may not have been included
in the early enrollment estimates but account for almost 1/7 of the increased enrollment since the
early estimate in January 2014.
What can be learned from automatic enrollments?
In early 2014, with approval of the Centers for Medicare & Medicaid Services (CMS), Vermont
automatically enrolled 33,549 plans from VHAP and Catamount into Medicaid. 6,7

Waiver approval received in a letter from CMS to Douglas Racine, Secretary of the Vermont Agency of Human
Services, dated February 27, 2014.
VT LEG #318187 v.1

Analysis of tax filers in 2013 showed that about 60 percent of the group that appeared eligible for
VHAP in December 2013 would qualify for Medicaid post-ACA, but actual enrollment in October 2014
was 85 percent of December 2013 VHAP enrollees (DVHA).
The CHAP population is harder to pin down. It seems reasonable that between 5 percent and 20
percent of CHAP enrollees might qualify for Medicaid post-ACA, but actual enrollment in October
2014 was about 34 percent (DVHA).
Higher automatic enrollments than seem reasonable today may be explained by the process of
automatic enrollment in early 2014 and the lack of redeterminations since then. Under Vermonts
transition plan approved by CMS, adults were considered eligible for automatic enrollment in
Medicaid based on the pre-ACA definition of income using 138 percent of FPL. The understanding
was that the post-ACA definition of income would be applicable when individuals came up for
renewal in 2014. No renewals or redeterminations occurred until late 2015, however, owing to
problems with Vermont Health Connect. The delay in redeterminations was approved in a waiver
from CMS such that redeterminations should be completed by February 2016 for legacy enrollees
who had received Medicaid prior to January 2014.8 That same waiver required that redeterminations
should be completed by November 2016 for Medicaid enrollees who first entered through Vermont
Health Connect using the new income definition. The State is acting within those guidelines.
Using tax filers only and rough measures of income eligibility suggests up to 1/4 of the transferred
VHAP adults (about 7,300 adults) and a large portion of the transferred Catamount adults (between
2,400 and 4,400 adults) would not have been eligible for Medicaid under the post-ACA definition of
income. If true, the number of adults enrolled in Medicaid today could be reduced significantly
through redeterminations. The difference would explain up to 1/3 of the increased enrollments since
the early estimate in January 2014.
Many of the former VHAP/Catamount adults no longer eligible for Medicaid could be eligible for
Vermont Premium Assistance at lower cost, of course.
Important caveats and considerations
a. Adults who are tax filers represent about 92 percent of resident adults ages 23 to 64; we dont
know whether non-filers have high incomes or low incomes, and we dont know the proportion of
Medicaid enrollees who are non-filers. Adults ages 18 to 22 were not included in the tax analysis.
b. Available tax information for filers is incomplete, both pre- and post-ACA.
7

Vermont Health Connect Update, Commissioner Mark Larson, April 29, 2014.
http://legislature.vermont.gov/assets/Documents/2014/WorkGroups/House%20Health%20Care/Vermont%20Health%
20Benefit%20Exchange/W~Mark%20Larson~Vermont%20Health%20Connect%20Update%204-29-14~4-29-2014.pdf.
8
A waiver from CMS dated November 13, 2015 allowed Vermont to delay income verification until 2016.

VT LEG #318187 v.1

c. The Medicaid resource test was dropped for people who qualify based on the new income rules.
VHAP and Catamount Health did not have a resource test, consistent with the New Adult rules.
d. The test regarding recency of health insurance no longer applies. Eligibility for VHAP and CHAP
required no health insurance coverage during the previous 12 months or a qualifying event such as
the loss of insurance through a spouse. That type of test is inconsistent with the federal insurance
mandate.
e. A complex and tedious eligibility process has become much easier. VHAP carefully checked
monthly income and dropped or added enrollees monthly. The ACA established a preference for
annual income redeterminations in order to reduce churn and ensure continuity of coverage.
f. ACA outreach and publicity increased awareness of Medicaid for many people. Vermont Health
Connect put income-eligible applicants in Medicaid as the default.
g. The new federal mandate for health insurance matters for a number of people.
h. Family size reflected in tax filings may differ from family size claimed for Medicaid eligibility.
Key acronyms and terms
ACA = the Affordable Care Act passed by Congress, effective in January 2014
BAA = Budget Adjustment Act, the budget act passed in the middle of a state fiscal year
CHAP = Catamount Health Premium Assistance Program
FPL = the federal poverty level, varies by family size, used to determine eligibility for Medicaid
MAGI = modified adjusted gross income, the new definition of income used for Medicaid eligibility
under the ACA
SFY = State fiscal year for Vermont, July 1 to June 30
VHAP = Vermont Health Assistance Program
Full Medicaid coverage types:
General Adults = adults with dependent children or pregnant women who qualify for full Medicaid
coverage based on a number of eligibility categories as well as income tests
New Adults = adults ages 18 to 64 who qualify for full Medicaid coverage based on their Modified
Adjusted Gross Income (MAGI)
Note: Any comments from the Administration will be posted on the JFO website.

VT LEG #318187 v.1